Huberman LabBringing Extinct Species Back to Life | Dr. Beth Shapiro
CHAPTERS
- 0:00 – 3:41
Why de-extinction work also helps prevent extinctions (dire wolf example)
Beth Shapiro opens with a concrete description of how Colossal approached the dire wolf: sequencing fossil genomes, selecting key genetic edits, and engineering those edits into a gray wolf genome. She immediately frames de-extinction as a dual-purpose effort—building the same tools needed to protect living species from disappearing.
- •Dire wolf approach: sequence fossils → identify trait-linked variants → make targeted edits in a living relative
- •De-extinction and conservation share the same core technology stack
- •Public excitement about extinct animals can drive funding and engagement for conservation tools
- 3:41 – 9:25
What a “species” is: taxonomy vs. breeding vs. genetics (bison & naming pitfalls)
Shapiro explains that “species” is a human-made category used to communicate, not something biology inherently obeys. Using bison/buffalo and Linnaean taxonomy, she contrasts naming systems with different species concepts, especially the biological species concept (interbreeding) versus genetic similarity thresholds.
- •Species categories are conceptual tools; biology doesn’t follow our boxes
- •Taxonomy (Linnaeus) organizes evolutionary relationships but doesn’t settle species boundaries
- •Biological species concept: fertile offspring implies “same species,” but it’s not universal
- •Genetic and geographic species concepts can be valid depending on the goal (e.g., conservation)
- 9:25 – 17:38
Neanderthals, Denisovans, and what ancient DNA revealed about human lineages
The conversation shifts to paleoanthropology and how ancient DNA transformed debates that were once based on sparse, contentious fossils. Shapiro explains how sequencing bones (e.g., Denisovan finger bone) reveals ancestry, confirms multiple contemporaneous human lineages, and documents interbreeding with Homo sapiens.
- •Ancient DNA reduces reliance on fragmentary fossil interpretation alone
- •Discovery of Denisovans from minimal remains via DNA sequencing
- •Humans met and interbred with Neanderthals after leaving Africa
- •Species boundaries blur when interbreeding is documented
- 17:38 – 22:58
Hybrid ancestry in nature: polar bears × brown bears and why gene flow can be one-way
Shapiro uses polar/brown bear hybridization to illustrate that interbreeding can persist even after long divergence times. She explains asymmetric gene flow (polar bear ancestry in brown bears but not vice versa) and how ecological selection (e.g., fur color affecting seal hunting) filters which hybrids survive.
- •Polar and brown bears diverged ~500k years ago yet still interbreed
- •Hybridization can be directionally biased (maternal/paternal constraints)
- •Natural selection can prevent hybrids from surviving in one niche (polar bear hunting)
- •Ancient DNA can reveal long-term patterns of admixture
- 22:58 – 32:04
What Neanderthal DNA does in modern humans: immunity, disease risk, red hair, eye color
Shapiro explains that each person’s 2–5% Neanderthal DNA is in different genome segments, and pooled across humans it covers most of the Neanderthal genome. She highlights cases where Neanderthal-derived variants rose in frequency due to past benefits—especially immune traits—while also discussing popular traits like red hair and speculation around eye color selection.
- •Neanderthal segments differ across people; collectively humans retain most Neanderthal sequence
- •Signals of selection: when Neanderthal variants exceed baseline frequency, they likely conferred advantages
- •Immune-related variants: examples include COVID risk/protection alleles
- •MC1R and evidence supporting red-hair variants in Neanderthals (and parallels in mammoths)
- •Eye color discussion: likely sexual selection; ancestral state is dark eyes
- 32:04 – 39:40
How to choose a de-extinction target: technical limits, ecology, and avoiding “re-extinction”
Shapiro outlines a pragmatic framework for selecting species to revive: DNA availability, knowledge of extinction causes, and whether the ecological niche still exists. She explains why dinosaurs are off the table (no recoverable DNA), why mammoths are feasible (oldest recovered DNA), and why ecosystem function—not nostalgia—should guide priorities.
- •Target selection criteria: technical feasibility, ethics, ecology, and social impact
- •Dinosaurs: too old; fossils are rock and DNA degrades rapidly after death
- •DNA degradation drivers: UV damage, freeze-thaw, microbial decay
- •Need to understand extinction causes to avoid immediate failure after revival
- •Evaluate ecological role and whether reintroduction would strengthen ecosystem resilience
- 39:40 – 50:19
Why mammoths and dodos: building a cross-species synthetic biology ‘toolkit’ (and bird constraints)
Shapiro argues that de-extinction is a platform: multiplex genome engineering, cell technologies, and phenotype mapping that can transfer directly to conservation. She explains why Colossal added a bird program and why birds are harder than mammals, motivating dodo work as both a technical challenge and a public-engagement catalyst.
- •De-extinction tool stack is directly reusable for conservation of living species
- •Birds are among the most endangered, but cannot be cloned via standard mammalian SCNT methods
- •Dodo chosen partly for engagement/awe and partly to advance bird-specific reproductive/genome tools
- •Bird genomics is tricky: microchromosomes and unusual germline-restricted chromosomes
- 50:19 – 56:42
What de-extinction actually builds: mammoth = edited elephant, not a “perfect clone” (Jurassic Park myths)
Shapiro clarifies that Colossal is not reconstructing an extinct genome from scratch or filling gaps with unrelated DNA à la Jurassic Park. Instead, they leverage the near-identity between mammoth and Asian elephant genomes and focus edits on regions consistently different between mammoths and elephants to engineer cold-adapted traits into an elephant background.
- •Mammoth and Asian elephant genomes are ~99% similar; approach is targeted editing, not full reconstruction
- •Goal is functional: an elephant capable of living in mammoth-like habitats today and tomorrow
- •‘Genetic species’ gatekeeping can distract from ecological/function goals
- •Key strategy: identify mammoth-consensus differences vs. elephant and edit those loci
- 56:42 – 1:11:41
Dire wolves in practice: selecting safe edits, avoiding harmful byproducts, and raising wild animals
Shapiro details how Colossal chose 20 dire-wolf-associated edits, including size and coat traits, while prioritizing animal welfare and safety on a gray-wolf genetic background. She explains why they avoided albinism-associated pathways for coat color, introduces the existing animals (Romulus, Remus, Khaleesi), and describes behavioral differences between hand-reared wild canids and domestic dogs.
- •Dire wolf creation used 20 deliberate edits informed by fossil genomes
- •Safety-first editing: avoid variants that could cause albinism-related blindness/deafness
- •Use alternative edits known to be safe in living wolves/dogs to achieve similar phenotype
- •Current animals: Romulus, Remus (males), Khaleesi (female); not breeding due to relatedness
- •Hand-rearing affects behavior; they are still clearly wild and physically large
- 1:11:41 – 1:18:12
Ecology, conservation ethics, and the ‘doing nothing is also a decision’ argument
The discussion broadens to ecosystem effects of (re)introducing species, emphasizing resilience and trophic cascades (e.g., Yellowstone wolves). Shapiro argues conservation already involves active human management—vaccination, predator protection, translocations—and that rejecting new tools implicitly accepts accelerating biodiversity loss.
- •Ecosystem resilience can improve by restoring missing interactions (top predators, herbivores)
- •Yellowstone wolves example: trophic cascades affecting vegetation and river dynamics
- •Conservation is already interventionist: managing numbers, food, disease, predators
- •Risk assessment must compare intervention risks against the risk of inaction (continued extinctions)
- 1:18:12 – 1:26:33
Gene drives, mosquitoes, invasive species, and modeling ecosystem outcomes (digital twins)
Shapiro distinguishes between eradicating all mosquitoes and targeting specific disease-carrying species or traits, noting human-altered environments inflate mosquito impacts. She defends gene drives as powerful but controllable tools (limited generations, selection against drives) and discusses using ecological modeling and “digital twin” approaches to forecast interventions.
- •Not all mosquitoes transmit disease; population spikes often come from human-built habitats
- •Alternative goals: reduce disease transmission rather than kill mosquitoes outright
- •Gene drives can be designed with time limits and safety switches; selection can counteract them
- •Example invasive threat: cheatgrass increases wildfire risk and outcompetes native grasses
- •AI-enabled ecosystem modeling (‘digital twins’) could test interventions before deployment
- 1:26:33 – 1:40:07
Governance, public trust, and human genetics: advisory panels, IVF selection, CRISPR medicine, and artificial wombs
Shapiro explains Colossal’s governance approach: local advisory panels and stakeholder-led stewardship decisions (e.g., Tasmania; Māori-led moa project). The conversation then pivots to human genetic selection and editing—IVF screening, the controversial CRISPR babies case, and the near-term future of gene therapy—culminating in a discussion of artificial womb technology as both a scaling tool for mammoths and a potential medical breakthrough for human pregnancy complications.
- •Colossal uses local advisory panels and community stewardship frameworks for potential releases
- •Human genetics is already shaped by selection (mate choice, IVF embryo screening)
- •CRISPR babies controversy illustrates where society draws lines—and how those lines can shift
- •Real-world therapeutic success: bespoke CRISPR editing for severe pediatric disease
- •Artificial wombs: needed for scaling de-extinction ethically and may enable earlier cancer treatment or fetal surgery in humans
- 1:40:07 – 2:05:16
Genetic rescue case study: black-footed ferrets, cloning for diversity, and engineering disease resistance
Using black-footed ferrets, Shapiro illustrates how multiple interventions can combine: captive breeding, cloning to restore lost genetic diversity from archived tissues, and future gene editing to address plague susceptibility. She also discusses how extreme bottlenecks can sometimes purge deleterious alleles (channel island foxes), while still leaving populations vulnerable to new threats.
- •Black-footed ferrets collapsed due to prairie dog control efforts; rediscovered in Wyoming
- •Captive breeding succeeds but suffers from low founder diversity over time
- •Cloning from frozen zoo tissues can reintroduce otherwise-lost genetic variation
- •Major current threat is plague; potential for gene edits to confer resistance (like domestic ferrets)
- •Bottlenecks can purge harmful recessives (island foxes) but reduce resilience to novel disease
- 2:05:16 – 2:15:36
Beth Shapiro’s origin story: journalism to Rhodes Scholar to ancient DNA pioneer (and why mammoths, not dinosaurs)
Shapiro recounts shifting from broadcast journalism to science after an immersive geology/archeology field course changed the kinds of stories she wanted to tell. A Rhodes Scholarship took her to Oxford, where she joined an early ancient DNA lab—attracted by interdisciplinarity, novelty, and fieldwork in Siberia—then explains why the “dinosaur question” persists and why mammoths became the culturally iconic (and technically plausible) alternative.
- •Career pivot: journalism → science storytelling via geology/archeology experiences
- •Rhodes Scholarship path to Oxford; joining Alan Cooper’s early ancient DNA lab
- •Ancient DNA appealed as a new interdisciplinary way to reconstruct past ecosystems
- •Media fixation on dinosaurs is longstanding; DNA preservation makes them infeasible
- •Mammoths persist as the public’s favorite: recency, scale, cultural imagery, and available DNA
